From @franknoe on June 24, 2015 5:42
In msm.analysis.(dense/sparse).decomposition, we have successfully used the reversible-option in eigenvalues in order to quickly compute eigenvalues of reversible matrices. Extend this functionality to eigenvector and RDL decomposition. The right/left eigenvectors can be computed from the eigenvectors of the symmetrized matrix by multiplying them by 1/sqrt(pi) or sqrt(pi), respectively.
Copied from original issue: markovmodel/PyEMMA#374
From @franknoe on June 24, 2015 5:42
In msm.analysis.(dense/sparse).decomposition, we have successfully used the reversible-option in eigenvalues in order to quickly compute eigenvalues of reversible matrices. Extend this functionality to eigenvector and RDL decomposition. The right/left eigenvectors can be computed from the eigenvectors of the symmetrized matrix by multiplying them by 1/sqrt(pi) or sqrt(pi), respectively.
Copied from original issue: markovmodel/PyEMMA#374